Industrial facilities of all kinds are trying to get a handle on their energy use in the interest of making their operations more efficient. The task makes it important to figure out how much power specific loads and circuits consume. Thus facilities are putting a priority on energy sub-metering -- quantifying energy consumption on each individual circuit feeding into the main breaker panels. Sub-metering can help identify the energy waste caused by defective appliances or inefficient electrical loads.
Moreover, sub-metering often helps detect abnormal behaviors faster than traditional sensors such as temperature, pressure, vibration, and so forth. An analysis of how these electrical parameters change over time even helps anticipate failures and plan predictive maintenance.
One device that is increasingly useful for sub-metering schemes is the contactless self-powered split core current transformer. These transducers can simply snap over a live conductor, without the need to screw or weld on complex brackets. Even better, they can install in electrical control panels, thus avoiding the need to add wiring or otherwise disturb electrical panels for measurement purposes.
Recently, these transformers have become widely used thanks to new types of ferrite that are more sensitive to magnetic fields and which are less expensive than older material. In addition, Rogowski coils, another contactless current sensing technology, recently have seen improvements that let these small, light and flexible sensors sense high currents. Better manufacturing processes have reduced the cost of these devices and made them less sensitive to how they are positioned around the current-carrying cable, until now one of the main problems of this promising technology.
Of course, transformers aren’t the only technology available for measuring current levels. Low-current applications may use a shunt. Outputs from these low-resistance devices are typically in the 50 to 100 mV range. The problem with shunts as measuring devices follows from I2R: Dissipated power rises rapidly as current increases. Shunts themselves are extremely accurate and have the best bandwidth of any current sensing method. But the subsequent isolators, amplifiers and analog-to-digital converters in the measurement stream will have a significant impact on the uncertainty of the measurement.
Hall-Effect transducers are another means of measuring current. They have voltage or current outputs and provide galvanic isolation. The isolation is inherent thanks to the Hall effect’s use of magnetic coupling. The transducer output is generally at levels that can feed into an a/d stage. Power consumption is significantly less than a shunt, especially at higher current ranges.
Shunts are competitive at current levels below about 10 A and where there is a need for high bandwidth. Hall effect devices work below about 150 kHz and across the spectrum of currents. New developments in ASIC technology are making low-cost open-loop Hall cell solutions competitive.
Other qualities such as power consumption, offset drift or gain drift over temperature can be important when evaluating current sensor technologies. In particular, uncertainty at temperature is often overlooked. Most applications do not operate at 25°C. Additionally, not all manufacturers use the same method to arrive at the accuracies stated on their respective datasheets. Several variables can affect pricing but the open-Loop Hall-effect devices with or without an ASIC are generally considered to be the least expensive solution.
Finally, flux-gate sensors are sometimes used to measure currents in pulse-width-modulated power supplies. They typically consist of a saturable magnetic core toroid with windings around one section of its radius. They detect the saturation state of a magnetic circuit. They also have an excitation winding to which a square voltage waveform is applied. In the absence of a primary current flowing through the center of the toroid, the average sensed current is zero (because the excitation signal is a symmetrical square wave). The flow of primary current moves the average current away from zero to a degree that can be measured.
Traditional donut-shaped solid-core current sensors are based on the principle of a transformer, where a core magnetically links primary and secondary windings. Current flowing through the conductor in the center of the core induces a magnetic field in the core. The magnetic field generates a current in the secondary windings proportionate to the primary current divided by the number of secondary winding turns. These basic current transformers are designed to measure sinusoidal alternating currents in the typical 50/60 Hz range. This well known technology is affordable because it uses common materials and processes.
Of course, the problem with solid-core current transformers is that their use requires shutting down power, disconnecting cables, and routing the wire of interest through the core. Split-core current transformers solve this issue but can be more expensive and less accurate than a similar solid-core transformer. It is thus important to understand differences among the various technologies available.
One source of inaccuracy in split-core current transformers is the fact that the magnetic core is in two distinct parts. Specifically, inaccuracy mainly comes from the imperfect contact between the two halves and the distribution of the secondary windings only around the two core halves, not uniformly around the entire core. Thus the contact surfaces must be quite flat and there must be enough pressure between the two to keep them in intimate contact. The case housing the current sensor generally incorporates flexible parts or materials and/or hinges to produce sufficient compression as well as a reliable opening mechanism.
FeSi has been widely used for core material mainly for its low price. But it exhibits a poor linearity (especially at low currents) and a large phase shift between the actual and the recorded waveforms. These qualities restrict its use to inexpensive settings for measuring rather large currents where high accuracy isn’t a priority.
Many applications only need a rough estimate of power consumption and aren’t hurt by the assumption that voltage waveform relationships are fixed. In such cases the high phase shift is not a big issue. A typical application is in monitoring branch currents coming out of control panels so a system can detect when circuits overload and either generate an alarm or rebalance loads.
Another disadvantage of FeSi current transformers is that they are large and heavy. So they may not work in environments with limited space.
FeNi has been the best material for split-core current transformers for a long time. It performs well but has been pricey. It offers a good alternative to the FeSi material when accuracy and phase shift are important, or for transformers that must measure small currents.
FeNi current transformers have other limitations. They are about as bulky as FeSi current transformers and exhibit pronounced linearity and drift, mainly because of the air gaps inherent in the split core architecture.
Ferrites have been well known core materials, but they have tended to magnetically saturate at relatively low levels of electric field strength. This property has kept them from being used at frequencies as low as 50/60Hz. However, recent developments have revolutionized the qualities of ferrite at these frequencies. New types of ferrites have significantly better permeability and can be implemented in 50/60 Hz current transformers as a substitute for FeSi or FeNi cores, despite the low magnetic saturation level.
Split-core current transformers incorporating the new types of ferrite can accurately measure ac signals in an extended frequency range that includes the 50/60 Hz domain. They take advantage of intrinsic ferrite qualities, providing high accuracy and excellent linearity even at low current levels. They also exhibit a particularly low phase-shift between input and output currents, essential for accurate measurement of true active power or energy. The hard and dense core helps minimize the gap between the two pieces and is virtually insensitive to ageing and temperature changes, unlike materials like FeSi or FeNi. Finally, all the ferrite qualities are available at a low cost.
It takes a bigger ferrite core to measure high currents. Unfortunately, bigger cores are rare for now because of fabrication limitations. So today, FeNi transformers or the Rogowski coil technology are more appropriate for high currents.
High-permeability ferrite material is relatively hard; consider it a ceramic. So it can be machined to a high degree, providing air gaps down to a few microns that are stable for many years. Laminated materials such as FeSi or FeNi do not allow air gaps smaller than 20 or 30 microns, and these are more sensitive to ageing and temperature changes. Add the small air gaps to the better linearity of the ferrite at low magnetic excitation (i.e. for low current), and the ferrite offers a better performance than FeNi-80%, and at a lower cost.
The phase shift for the ferrite is half that of FeNi cores. The reduced air gap of the ferrite core also brings a more accurate transfer ratio (primary turns to secondary turns).
A Rogowski Coil consists of a helical coil of wire wound in a relatively tight radius with the lead from one end returning through the center of the coil to the other end, so both terminals are at the same end of the coil. The whole assembly is then wrapped around the straight conductor of interest. The voltage is induced in the coil is proportional to the rate of change (derivative) of current in the straight conductor. Thus, the output of the Rogowski coil usually goes to an integrator circuit to yield an output signal proportional to the current. In other words, the Rogowski coil forms an accurate and linear current sensor at the price of additional electronics and calibration. The length of the coil is selected according to the amount of current to be measured.
A Rogowski coil has a lower inductance than current transformers because it has no magnetic core material and consequently exhibits a better frequency response. It is also highly linear, even with high primary currents, because it has no iron core that may saturate. This kind of sensor is thus particularly well adapted to power measurement systems that can see high or fast-changing currents. For measuring high currents, it has the additional advantages of small size and easy installation.
The performance of a Rogowski coil depends on how well it is made; its windings must be equally spaced to provide high immunity to electromagnetic interference. Also critical is the closing point: It induces a discontinuity in the coil, creating some sensitivity to external conductors as well as to the position of the measured conductor within the loop. The locking or clamping system should ensure the coil extremities stay at a precise and reproducible position. It should also maintain a high symmetry while one extremity connects to the output cable.
Some new technologies have recently appeared in this area, with special mechanical and electrical qualities that promote accuracy and immunity even while the primary cable sits in a variety of positions. While the error arising from primary cable position was typically not better than ± 3% in the 50/60 Hz frequency domain, it has been reduced to less than ± 0.5% on some of the latest Rogowski coil sensors.
Navigating sensor specs
Current sensors can be judged according to six key qualities:
Accuracy - The accuracy of a power calculation obviously depends on the accuracy of the current sensors. A class 1 power meter will need current sensors with much better than 1% accuracy, which would generally involve expensive materials and sophisticated manufacturing. An alternative is to calibrate the power meter for each single sensor it uses. Accounting for the specific qualities of each sensor lets a power meter operate in its most precise operating mode.
Drift - The drift of a sensor relates to how a reading varies over time independent of the initial system calibration. Changes in the ambient humidity and temperature, component aging, and so forth, all affect drift. A low drift level - meaning that the sensor has is highly immune to such constraints - is important when building high-performance, stable and reliable power meters.
Linearity - The linearity of the sensor refers to the stability of its qualities within the full operating mode. A high linearity is essential for accurately measuring primary currents that span a wide range, especially those at low levels. Several technologies only work well over a limited measuring range, generally limiting their use to rather high or low currents.
Phase shift - The accuracy of the true active power or energy calculation relates not only to how accurately and linearly the ac current and voltage sensors record amplitude; the phase shift that may arise between the measurement of these correlated values also plays a role. The phase shift should of course be as low as possible.
Integration - Being passive devices, current transformers need no wiring other than a two-wire output connection to the main power monitor unit. High-accuracy power meters require specific calibration for each sensor. They may have low-current outputs that are safer than traditional 1 or 5-A signals which can be accessed while the system operates. Current outputs are also almost insensitive to interference and should be preferred to voltage outputs when long cables connect the sensors to the power meter.
Price – Sensor price is important, of course, especially when the task calls for three accurate current sensors measuring three-phase power. However, the price of the current sensor isn’t the only consideration. Installation and maintenance costs can be significant, particularly when wiring must be powered down and unhooked while technicians add solid-core sensors. This is why split-core sensors can reduce costs as a whole.
LEM USA Inc., Milwaukee, Wis., www.lemusa.com